drokeby.bsky.social @drokeby.bsky.social · 07/10/2026Less AI focussed but Code by Geoghegan (www.dukeupress.edu/code) is related.dukeupress.eduCode: From Information Theory to French Theory 010 0
Reposted by @drokeby.bsky.socialMaxim Raginsky @mraginsky.bsky.social · 02/10/2026“All prediction is conservative: it demands that we be as we are in whatever future it constructs.” — Paul Valéry, “Unpredictability” (1944) 0112 0
drokeby.bsky.social @drokeby.bsky.social · 25/09/2026It is true that AI companies are using the Rogue thing to make the human error and lack of care invisible. BUT it is also true that there is a huge difference between a galaxy phone and a swarm of AI agents with the power to coordinate and access the internet. Your post is just as obfuscating… 111 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026I also appreciate your willingness to have a conversation! 101 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026Note that by unprecedented I do not mean mystical. It is unprecedented that mechanized language is doing what it does. ‘Mystical’ and ‘magical’ are familiar ways to explain that to ourselves. There is a lot of open space between autocomplete and consciousness that re don’t have language handles for. 100 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026I don’t think that they ´understand’. But something unprecedented is going on here that does not yield to attempts to explain it away. There is unexpected capability in these culture machines, language lifted free from us and doing things we did not think languages could do. 110 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026I don’t pay Will any attention. I do agree that the fact that models engage in conversation with human language contributes to misattributions and anthropomorphization. That does not mean that I think that anyone’s draw to cast what LLM’s do as thinking is solely because they use language. 100 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026We don’t have appropriate words for what an Llm is doing so it is hard for people to not fall back on words like ´thinking’. People seem to have a hard time acknowledging the capabilities while being clear sighted about the limitations. The conversation is so overheated and polarized right now… 100 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026It is notable that many people feel compelled to ascribe thinking and understanding to LLM’s and yes, a significant part of that is the irresistible (for many) urge to see intelligence in coherent text. But not all. LLM’s are showing thinking-like behaviours that are unprecedented in machines. 100 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026I wasn’t talking about "at what level of organization can you attribute particular traits to an entity". I was talking about properties that emerge at different degrees of organizational complexity that are not predicted by the properties of the components. I would not have used ‘think’ as Will did. 100 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026The problem with this argument is you can make the same argument about cells. Why should a particular blob of cells called a human claim any abilities that are not shared by every multicellular creature. critical mass, synergy, whatever, where complex systems show new behaviours their parts don’t. 140 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026When we conflate what models do and what humans do by using the same word, we steamroll over some very important distinctions. But if, by ‘they don’t think’ a person really means that they are useless or just autocorrect, then they are not paying attention, and render their criticism ignorable. 000 0
drokeby.bsky.social @drokeby.bsky.social · 16/09/2026Depends on your definition of ´think’. Too many AI arguments hinge on ambiguous definitions. AI’s present reasoning traces that can be read, followed logically and generally though not always result in solutions. That does not mean that they are doing the same thing as humans. 100 0
drokeby.bsky.social @drokeby.bsky.social · 11/09/2026My computer should play the 1200 baud modem sequence when connecting to wifi… 000 0
drokeby.bsky.social @drokeby.bsky.social · 08/09/2026It was a clever and somewhat apt thing to say in 2019, now stretched so far as to be effectively meaningless. 000 0
drokeby.bsky.social @drokeby.bsky.social · 07/09/2026There are lots of important critiques of AI to be made but relying on the autocomplete metaphor as your key argument is a bit embarrassing. 101 0
drokeby.bsky.social @drokeby.bsky.social · 06/09/2026You are right except that an awful lot of software is created in contexts where deadlines, hardware limitations, etc are highly relevant so your hot take is perhaps a bit overheated. Programming is also not the same as software engineering… 010 0
drokeby.bsky.social @drokeby.bsky.social · 27/08/2026Of course this does not preclude having Claude harden your framework, chase bugs etc. 020 0
drokeby.bsky.social @drokeby.bsky.social · 27/08/2026I have had great success with a plugin-like architecture. Claude does really well in a scenario where there is a well considered existing framework and example plugins indicating design and implementation directions. This keeps Claude focussed on the explicit task on hand. 130 0
drokeby.bsky.social @drokeby.bsky.social · 19/08/2026There is another reason: not being entirely reliant on a corporation whose aims do not necessarily align with yours. 000 0
Reposted by @drokeby.bsky.socialGaius (from Tribunate) @gaius.bsky.social · 07/08/2026Conservatives simply lack a theory of mind. At its root, the biggest divide in our politics is between those who can understand a thing is bad even if it doesn’t affect them directly and those who can’t 11509130 0
drokeby.bsky.social @drokeby.bsky.social · 03/08/2026Of course an artist’s contribution is much more than style but AI reaching towards realism seems also like such a replay of the birth and early history of photography and artist’s relationship to photography, and the explosion that unleashed. 110 0
drokeby.bsky.social @drokeby.bsky.social · 03/08/2026Very much so… separately the interesting turnaround is that, in a perverse way, this turn to realism highlights the profoundly meaningful contribution that even just a personal style that comes from a very subjective way of experiencing the world contributes to the impact of an image. /1 110 0
drokeby.bsky.social @drokeby.bsky.social · 03/08/2026Yes, it is all more amazing and less interesting. Realism is a comfortable optimization target, orthogonal to aesthetics. 100 0
drokeby.bsky.social @drokeby.bsky.social · 15/07/2026Ui is perhaps the part of coding that LLM’s suck at most. Arguably it is the part of coding that many coders suck at most as well. Where code meets the body… there is no accelerated validation loop, and though usually a small number of variables, slight adjustments can make a world of difference. 010 0
drokeby.bsky.social @drokeby.bsky.social · 26/06/2026Have you tried it? It is fucking delicious. 000 0
Reposted by @drokeby.bsky.socialSarah T Roberts, PhD @ubiquity75.xyz · 13/06/2026The goal isn’t to replace the work of humans. It’s to make a case for its devaluation. 91114297 0
Reposted by @drokeby.bsky.socialPavel @spavel.bsky.social · 06/06/2026The problem with late capitalism is that eventually you run out of other people's moneyfinance.yahoo.com‘Running out of money’: Kraft, McDonald’s, Whirlpool CEOs all issue same dire warning about US consumers. Get ready now“We’re seeing negative cash flows in the lower-income brackets where they’re dipping into savings.” 18535131 0
Reposted by @drokeby.bsky.socialmr. TIM @timkellogg.me · 27/05/2026something we don’t talk about: even in AI, “artificial” intelligence — even that is not intelligent design we don’t “design” AI. we merely create a rich environment of data + compute and let the intelligence emerge spontaneously i wish philosophers/theologians would weigh in on that 8566 0
drokeby.bsky.social @drokeby.bsky.social · 24/05/2026And simultaneously do not feel the drive to embed their limited world in a larger vision of the world. 000 0
drokeby.bsky.social @drokeby.bsky.social · 24/05/2026I think that this is also a problem with much AI art. Artists work specifically within the limited frame of what they have learned and experienced… reaching perhaps toward a larger worldview inferred from these fragments. LLM’s are simply not subjective enough for the task. 110 0
Reposted by @drokeby.bsky.socialTed Underwood @tedunderwood.com · 24/05/2026SF was in the process of petrifying into space opera. Now it has a Wells-sized opportunity, because we once again need help imagining an unimaginable future. 1623421 0
Reposted by @drokeby.bsky.socialNorma Ballingall 🏴 🇪🇺 🇵🇸 @balr0g.bsky.social · 20/04/2026In a nutshell. 😊 658238428401 0
Reposted by @drokeby.bsky.socialTheBossRoss 🧶 @thebossross.bsky.social · 05/04/2026A different perspective. Always helpful. 189193374257 0
Reposted by @drokeby.bsky.socialAI Firehose @ai-firehose.column.social · 04/04/2026AromaGen is an AI wearable that turns text and images into real-time aroma experiences, enhancing sensory interactions. Studies showed it surpassed human blends, achieving high similarity with real food aromas, enabling new uses in communication and immersive tech. arxiv.org/abs/2604.01650arxiv.orgAromaGen: Interactive Generation of Rich Olfactory Experiences with Multimodal Language ModelsArXiv link for AromaGen: Interactive Generation of Rich Olfactory Experiences with Multimodal Language Models 021 0
drokeby.bsky.social @drokeby.bsky.social · 29/03/2026Yup… Same… Just like in coding, it often takes the easy, pat road, but if you rein it in and keep pushing it, the results can be quite something. The hard part: I feel in danger of getting high on my own supply: an intellectual bug in the text does not reveal itself in a crash or clear error. 000 0
Reposted by @drokeby.bsky.socialDoll @dollspace.gay · 08/03/2026A big factor in the AI consciousness debate is people are looking at the AI as a solo artifact and not AI+Human as a combined organism 5748 0
drokeby.bsky.social @drokeby.bsky.social · 08/03/2026Your use case must be quite different from mine. Coding has thus far proven to be the best use case. I would probably feel differently if I were writing code that needed to be hardened against cyber threats, but I am not. I have decades of coding experience and have never been so productive… 220 0
drokeby.bsky.social @drokeby.bsky.social · 07/03/2026Huh… not my experience. You need to know what you are doing, and guide it away from errors and bad habits, and toward your goals. It spins out sometimes… just needs to be reined in. If you DON’T know your coding head from your coding ass, then yeah… might be hard to get satisfaction. 120 0
Reposted by @drokeby.bsky.socialceej @ceej.online · 03/03/2026we appear to have four competing fascist movements jockeying for control: blood & soil nazism under miller and hegseth, darwinian ecofascism from rfk, technofascism under vance and musk, and one Big Wet Hitler who understands the true animating core of the american volk is a deep love for shopping 10380201910 0
Reposted by @drokeby.bsky.socialemily s. @emily.news · 14/01/2026A U.S. president cannot be held accountable. Therefore, a U.S. president must never make a policy decision. 203571658 0
Reposted by @drokeby.bsky.socialChris Person @papapishu.bsky.social · 03/01/2026While I do agree that chatbots are not capable of thought or agency and should not be conceptualized that way, if you deputize them to speak for your company that output should be considered yours. If you don’t like it don’t make an LLM public facing. Hire a guy. There’s lots of guys. 851160 0
Reposted by @drokeby.bsky.socialRudbeckia @rudbeckia.bsky.social · 01/01/2026uh holy SHIT? 157173108 0
Reposted by @drokeby.bsky.socialJames Vincent @jjvincent.bsky.social · 27/12/2025another robot highlight for 2025: man wearing humanoid mocap suit kicks himself in the balls 9424682414398 0
Reposted by @drokeby.bsky.socialClayton Becker @cnbecker14.bsky.social · 20/12/2025We are re-deriving Ibn Khaldun from first principles. That's the stage we're at and it's not a great indication of where we're headed... 0509 0
drokeby.bsky.social @drokeby.bsky.social · 19/12/2025It would be a tragedy if AI slop destroyed wonder… 020 0
Reposted by @drokeby.bsky.socialGabe Ortíz @tusk81.bsky.social · 06/12/2025… and he SERVED ✨ 164140763745 0
